Former Google CEO Eric Schmidt estimates AI infrastructure at $50 billion per gigawatt, requiring half a trillion dollars for 10 gigawatts.
Former Google CEO Eric Schmidt identified capital, not energy, as the primary constraint on AI growth, estimating $50 billion per gigawatt of compute capacity. At this rate, scaling to 10 gigawatts would cost approximately $500 billion, a sum few entities can finance.
Schmidt noted China’s ability to mobilize such capital but expressed uncertainty about its current investments. The U.S. capital markets’ borrowing capacity provides a structural advantage, while Europe lacks the scale to fund AI infrastructure at this level.
The financial barrier could surpass energy constraints as the dominant limit on AI expansion, reshaping investment priorities for governments and corporations.